video buffering

I have an Internet connection that nominally runs at 100Mb/s. Most Web
pages load very quickly, and I sometimes even see much higher speeds, as
when I'm downloading a Win 10 upgrade image, where I see rates of 220Mb/s.
So my question is: why do so many video files stall so frequently? I
understand that is too vague a question to merit a good answer, but I
wonder if there are some obvious network settings I should try tweaking.
Where are the bottlenecks? Network Properties for the (hard-wired)
connection include some items that I don't understand (so I've left the
defaults alone). I imagine my router might be the culprit so wonder about
settings for it too.

On Mon, 14 Aug 2017 22:43:59 -0400 "Wolf K" wrote
in article
Every now and then, some (un)lucky organisation discovers that
a sudden surge in demand for its products or services overwhelms its
server(s).


I've seen that when there is something "breaking" that likely causes
servers to overload. What I see often is that a video will begin playing
and then stall partway through, often for a very long time. It's the very
long time that doesn't make sense to me.

Jason wrote:

I have an Internet connection that nominally runs at 100Mb/s. Most Web
pages load very quickly, and I sometimes even see much higher speeds, as
when I'm downloading a Win 10 upgrade image, where I see rates of 220Mb/s.
So my question is: why do so many video files stall so frequently? I
understand that is too vague a question to merit a good answer, but I
wonder if there are some obvious network settings I should try tweaking.
Where are the bottlenecks? Network Properties for the (hard-wired)
connection include some items that I don't understand (so I've left the
defaults alone). I imagine my router might be the culprit so wonder about
settings for it too.

No matter how much faster you have for downstream bandwidth, you cannot
get more than the server will give you. You might be sucking on a 4"
pipe but the site's server delivers through a straw. They will throttle
the connections to ensure ALL of them get some response. They don't all
their bandwidth to just one connection for super speed for a single user
while making all their other hundreds or thousands of other connections
sit there pending their turn. A piece of the pie for everyone, not the
whole pie for one.

You will find a lot of sites that will not deliver the bandwidth you are
capable of. Besides the above scenario to accomodate lots of
connections by many visitors, a site may be hosted at a less than
steller service tier with a webhoster. The webhoster requires their
customer to pay more for more bandwidth, so the customer may decide what
is appropriate for their visitors and what they can afford.

If you think your router is the problem, just bypass it. Connect your
desktop directly to the cable modem. That means all your other intranet
hosts won't have any network access (well, only your intranet for those
you leave connected to the router). The one test host you directly
connect to the cable modem will get all of your service tier's Internet
bandwidth (none of it is shared with your other hosts).

Depends on the web browser, too. I've seen long-lasting threads in the
Firefox newsgroup about site authors that get Google Chrome to smoothing
display their videos (with maybe just a pause to buffer at the
beginning) but find there are unacceptable jitter or lags when using
Firefox as the client to view their video.

On Tue, 15 Aug 2017 01:12:32 -0500 "VanguardLH" wrote in
article
If you think your router is the problem, just bypass it. Connect your
desktop directly to the cable modem.


I've tried that. It didn't make a difference that I could see. The machine
I usually use is wired directly to the router - no WiFi.

Wolf K wrote:

Jason wrote:

VanguardLH wrote:

If you think your router is the problem, just bypass it. Connect your
desktop directly to the cable modem.

I've tried that. It didn't make a difference that I could see. The machine
I usually use is wired directly to the router - no WiFi.

Then it's most likely server overload: watch the progress bar at the
bottom of the video, it should be grey some distance in front of the
timing dot. If so, try waiting a minute or so before you start playing
the video. By that time, there should be enough data in the buffer that
server-side slowdowns shouldn't be a problem.

As long as the local player doesn't catch up to the buffering.

No idea what "video files" the OP is trying to play inside of Firefox
(or using the Flash plug-in or using some Javascript player in the web
page). If it is indeed a file, typically you can download the file like
any other and play it locally without any concern about server load
(other than effect the time to download). More likely the video content
is streamed media. "Video files" is too vague to know if the OP is
actually retrieving files to view, if Flash is involved, what sites are
involved (i.e., where the OP tested), what add-ons aka extensions the OP
installed into his web browser, and if the OP tested with a different
web browser. As yet, which web browser the OP is using is still
unknown.

Might help to narrow the responses from shotgun troubleshooting
suggestions based on vague info by giving details. Also loading Windows
in its safe mode with networking and also loading the unidentified web
browser in its safe mode to eliminate any add-ons would eliminate
startup programs and add-ons as causes for the problem. Trying a new
(clean) profile in the web browser would eliminate any user-customized
settings, especially any that were changed in about:config (Firefox) or
the experimental settings (Chrome).

On Wed, 16 Aug 2017 14:41:01 -0500 "VanguardLH" wrote in
article
"Video files" is too vague to know if the OP is
actually retrieving files to view, if Flash is involved, what sites are
involved (i.e., where the OP tested), what add-ons aka extensions the OP
installed into his web browser, and if the OP tested with a different
web browser. As yet, which web browser the OP is using is still
unknown.

I am trying to view files in browsers - Chrome and Edge. Some are YouTube
videos, others are various..from news organizations etc. I'm not
downloading them for later viewing. I have tried your suggestion of
eliminating plugins and have tried Safe Mode too. Didn't make a difference
that I could see.

Jason wrote:

VanguardLH wrote:

"Video files" is too vague to know if the OP is
actually retrieving files to view, if Flash is involved, what sites are
involved (i.e., where the OP tested), what add-ons aka extensions the OP
installed into his web browser, and if the OP tested with a different
web browser. As yet, which web browser the OP is using is still
unknown.

I am trying to view files in browsers - Chrome and Edge. Some are YouTube
videos, others are various..from news organizations etc. I'm not
downloading them for later viewing. I have tried your suggestion of
eliminating plugins and have tried Safe Mode too. Didn't make a difference
that I could see.

I don't use Edge for anything so cannot help you with that one. In
Google Chrome (and Firefox), you can reset your current profile (or a
new one gets created). For Chrome:

https://support.google.com/chrome/answer/3296214?hl=en

Rather than blowing away your current profile or stuck in a new one by
default, you can create a new profile and manage which one to use. For
Google Chrome, see:

http://www.pcworld.com/article/20893...in-chrome.html

Not only does this eliminate add-ons as a cause (because those are
managed on a per-profile basis so they won't be in a new profile) but
also any user-changed configuration settings are set to their defaults
in a new profile; i.e., you start with a clean profile.

You might also want to see what else is consuming your bandwidth. Use
something like SysInternals' TCP View to see what connections there are
by which processes. You can look at the columns for Sent Bytes and
Received Bytes to see which ones have sucked up the most bandwidth. I
believe that a cumulative statistic: it begins when you load TCP View.
It doesn't show the load on bandwidth within, say, the last few seconds
to see what is chewing up bandwidth now.

My aunt was complaining about slow and irregular bandwidth (by noting
long stalls in the web browser). I found she had installed toolbars and
add-ons in her web browser (covertly installed) along with some
background processes for other spyware and malware. Once I got rid of
them, she was back to normal browsing.

On Wed, 16 Aug 2017 22:32:31 -0400, Jason
wrote:

On Wed, 16 Aug 2017 14:41:01 -0500 "VanguardLH" wrote in
article
"Video files" is too vague to know if the OP is
actually retrieving files to view, if Flash is involved, what sites are
involved (i.e., where the OP tested), what add-ons aka extensions the OP
installed into his web browser, and if the OP tested with a different
web browser. As yet, which web browser the OP is using is still
unknown.

I am trying to view files in browsers - Chrome and Edge. Some are YouTube
videos, others are various..from news organizations etc. I'm not
downloading them for later viewing. I have tried your suggestion of
eliminating plugins and have tried Safe Mode too. Didn't make a difference
that I could see.

Have you tried viewing video files with VLC media player?

Jason wrote:
On Wed, 16 Aug 2017 14:41:01 -0500 "VanguardLH" wrote in
article
"Video files" is too vague to know if the OP is
actually retrieving files to view, if Flash is involved, what sites are
involved (i.e., where the OP tested), what add-ons aka extensions the OP
installed into his web browser, and if the OP tested with a different
web browser. As yet, which web browser the OP is using is still
unknown.

I am trying to view files in browsers - Chrome and Edge. Some are YouTube
videos, others are various..from news organizations etc. I'm not
downloading them for later viewing. I have tried your suggestion of
eliminating plugins and have tried Safe Mode too. Didn't make a difference
that I could see.

For a "second opinion", I would try from a Linux LiveCD.
Run the provided browser there, and see if the same behavior
is present. You should be able to get Chromium in Linux
(using the Package Manager). Maybe a copy of Linux Mint
would be the least nuisance, as a test platform.
If both OSes do the same thing, it's probably not an
OS-level issue.

You could try running Wireshark, and see if there is any
sign the ISP is meddling with the connections. Net neutrality
isn't what it used to be. If the ISP drops packets using
a DPI box, then the connection rate may end up throttled.

It's probably not a persistent connection problem.

https://en.wikipedia.org/wiki/HTTP_p...ent_connection

Or a delay*bandwidth problem, as your file transfers work
at full rate. On an older OS, the TCP/IP window may not
be big enough, for such a fat pipe. The more modern the
OS, the more likely it handles this situation on its own.
Without using a tweaker.

Your regular file transfers running at full rate,
rules out a bunch of causes. I would be using Wireshark,
in the hopes of receiving a "hint" when looking at the trace.

*******

I was hoping to find an "off brand" source of video. This
one is actually protected with DRM, but seems pretty
smooth here. This plays good in SRWare Iron (a Chrome
clone).

http://www.dash-player.com/demo/drm-test-area/

The URL will redirect here.

https://bitmovin.com/mpeg-dash-hls-drm-test-player/

Does that work better than Youtube ?

Paul

On Thu, 17 Aug 2017 09:03:33 -0400 "Paul" wrote in
article
Your regular file transfers running at full rate,
rules out a bunch of causes. I would be using Wireshark,
in the hopes of receiving a "hint" when looking at the trace.


Thanks. I have that and will see what it reveals.
The video in your link played perfectly - very smooth.

Jason wrote:
On Thu, 17 Aug 2017 09:03:33 -0400 "Paul" wrote in
article
Your regular file transfers running at full rate,
rules out a bunch of causes. I would be using Wireshark,
in the hopes of receiving a "hint" when looking at the trace.


Thanks. I have that and will see what it reveals.
The video in your link played perfectly - very smooth.

Then this could well be a lack of net neutrality.

And using VPN tunnels as a means of testing for DPI
boxes, isn't a guaranteed win. Some ISPs will also
**** around with encrypted streams if they see them.
Really, nothing is sacred when it comes to cheating
customers. If they own a DPI box, it's being used
110% to interfere with traffic. And make the
triple-play Internet TV option look attractive
(as the DPI lets that pass).

https://en.wikipedia.org/wiki/Deep_packet_inspection

My previous ISP, had a mis-adjusted DPI box, which
splattered virtually all web connections at random,
with RST packets. What a mess... That causes connections
to drop. And the consumer has to do a lot of retries
to get anything done. Of course they didn't admit to it,
and there is no way sitting at my desk here, to
differentiate between legit RST packets (actually sent
by the server), versus RST packets (sent *both* ways
by the DPI box).

A DPI box is a MITM (man in the middle) attack.

Paul

On Thu, 17 Aug 2017 15:32:23 -0400, Paul wrote:

Jason wrote:
On Thu, 17 Aug 2017 09:03:33 -0400 "Paul" wrote in
article
Your regular file transfers running at full rate,
rules out a bunch of causes. I would be using Wireshark,
in the hopes of receiving a "hint" when looking at the trace.


Thanks. I have that and will see what it reveals.
The video in your link played perfectly - very smooth.

Then this could well be a lack of net neutrality.

And using VPN tunnels as a means of testing for DPI
boxes, isn't a guaranteed win. Some ISPs will also
**** around with encrypted streams if they see them.
Really, nothing is sacred when it comes to cheating
customers. If they own a DPI box, it's being used
110% to interfere with traffic. And make the
triple-play Internet TV option look attractive
(as the DPI lets that pass).

https://en.wikipedia.org/wiki/Deep_packet_inspection

My previous ISP, had a mis-adjusted DPI box, which
splattered virtually all web connections at random,
with RST packets. What a mess... That causes connections
to drop. And the consumer has to do a lot of retries
to get anything done. Of course they didn't admit to it,
and there is no way sitting at my desk here, to
differentiate between legit RST packets (actually sent
by the server), versus RST packets (sent *both* ways
by the DPI box).

A DPI box is a MITM (man in the middle) attack.

DPI doesn't have to mean MITM. Where I worked in about 2005, a large
nationwide ISP, we purchased a DPI 'solution', which we connected to
span ports on each switch. The DPI box got see all traffic, but it had
no way to interfere with that traffic. It wasn't in the path, it was
hanging off to the side.

Since then, I've set up and/configured DPI boxes at several ISPs and
corporations, always in that same passive mode.

OTOH, at one time Comcast was apparently doing exactly what you
described above. When they saw P2P traffic, they'd send a TCP RST in
both directions, so that both endpoints assumed that the other end had
aborted the connection. They got caught and supposedly stopped. I
haven't seen further reports in a very long time.

On 17/08/2017 23:11, Char Jackson wrote:
On Thu, 17 Aug 2017 15:32:23 -0400, Paul wrote:

Jason wrote:
On Thu, 17 Aug 2017 09:03:33 -0400 "Paul" wrote in
article
Your regular file transfers running at full rate,
rules out a bunch of causes. I would be using Wireshark,
in the hopes of receiving a "hint" when looking at the trace.


Thanks. I have that and will see what it reveals.
The video in your link played perfectly - very smooth.

Then this could well be a lack of net neutrality.

And using VPN tunnels as a means of testing for DPI
boxes, isn't a guaranteed win. Some ISPs will also
**** around with encrypted streams if they see them.
Really, nothing is sacred when it comes to cheating
customers. If they own a DPI box, it's being used
110% to interfere with traffic. And make the
triple-play Internet TV option look attractive
(as the DPI lets that pass).

https://en.wikipedia.org/wiki/Deep_packet_inspection

My previous ISP, had a mis-adjusted DPI box, which
splattered virtually all web connections at random,
with RST packets. What a mess... That causes connections
to drop. And the consumer has to do a lot of retries
to get anything done. Of course they didn't admit to it,
and there is no way sitting at my desk here, to
differentiate between legit RST packets (actually sent
by the server), versus RST packets (sent *both* ways
by the DPI box).

A DPI box is a MITM (man in the middle) attack.

DPI doesn't have to mean MITM. Where I worked in about 2005, a large
nationwide ISP, we purchased a DPI 'solution', which we connected to
span ports on each switch. The DPI box got see all traffic, but it had
no way to interfere with that traffic. It wasn't in the path, it was
hanging off to the side.

Since then, I've set up and/configured DPI boxes at several ISPs and
corporations, always in that same passive mode.

OTOH, at one time Comcast was apparently doing exactly what you
described above. When they saw P2P traffic, they'd send a TCP RST in
both directions, so that both endpoints assumed that the other end had
aborted the connection. They got caught and supposedly stopped. I
haven't seen further reports in a very long time.

I'm REALLY surprised that an important fellow like you doesn't have a
LinkedIn profile.

Why is that Char?

D.

David B. wrote:


I'm REALLY surprised that an important fellow like you doesn't have a
LinkedIn profile.

Why is that Char?

D.

Successful enough to get to keep some privacy maybe ?

Only people in dire need of promotion are on there.
Door to door salesmen. That sort of thing.

That's why Microsoft paid $26.4 billion, to buy the lot.
I wonder if they've worked out yet, how long that's going
to take to pay back ? If would take a lot of Office 365
subscriptions.

Paul

On 18/08/2017 00:02, Paul wrote:
David B. wrote:


I'm REALLY surprised that an important fellow like you doesn't have a
LinkedIn profile.

Why is that Char?

D.

Successful enough to get to keep some privacy maybe ?

I somehow doubt that. *I* think he's full of bull**** and frightened of
his own shadow! He's certainly not an honest netizen. Good guys don't
hide from reality. Truly!

Only people in dire need of promotion are on there.
Door to door salesmen. That sort of thing.

I'm confident that you are mistaken Paul. ;-)

Quote
*****

LinkedIn (/ˌliŋkt.ˈɪn/) is a business- and employment-oriented social
networking service that operates via websites and mobile apps. Founded
on December 28, 2002,[6] and launched on May 5, 2003,[7] it is mainly
used for professional networking, including employers posting jobs and
job seekers posting their CVs. As of 2015, most of the company's revenue
came from selling access to information about its members to recruiters
and sales professionals.[8] As of September 2016, LinkedIn had more than
467 million accounts, out of which more than 106 million are active.[5]
As of April 2017, LinkedIn had 500 million members in 200 countries.

LinkedIn allows members (both workers and employers) to create profiles
and "connections" to each other in an online social network which may
represent real-world professional relationships. Members can invite
anyone (whether an existing member or not) to become a connection. The
"gated-access approach" (where contact with any professional requires
either an existing relationship or an introduction through a contact of
theirs) is intended to build trust among the service's members. LinkedIn
participated in the EU's International Safe Harbor Privacy Principles.

The site has an Alexa Internet ranking as the 20th most popular website
(October 2016).

https://en.wikipedia.org/wiki/LinkedIn

That's why Microsoft paid $26.4 billion, to buy the lot.
I wonder if they've worked out yet, how long that's going
to take to pay back ? It would take a lot of Office 365
subscriptions.

Indeed!

FYI

From: Char Jackson
Newsgroups: alt.computer.workshop
Subject: video buffering
Message-ID:
NNTP-Posting-Date: Fri, 18 Aug 2017 04:54:40 UTC

No, you're not surprised, because we've discussed it before, more than
once. Check your dossier and you'll see that I don't have a LinkedIn
profile because there are stalkers like you who are only too eager to
abuse it for their own nefarious purposes.

=

Here's an item published by my protégé - I'm very proud of what he has
achieved. :-)

https://www.linkedin.com/pulse/fund-...than-wadsworth

--
"Do something wonderful, people may imitate it."
https://www.linkedin.com/in/boaterdave/